AI for Good Live Pitching Session #03
7:10PM Jun 26, 2020
Morning, Good afternoon or good evening. Welcome to the third starts pitching session with AI for Good innovation factory. My name is Andrea from ITU that International Telecommunication Union. And I have the privilege of facilitating this pitching session today. The ITU the International Telecommunication Union, is the United Nations specialized agency for icfes. And we are also the organizers of AI for Good Global Summit alongside the XPrize Foundation, co convened by Switzerland, and in partnership with 36 United Nations agencies and ECM. The goal of the summit is to identify practical AI applications that can advance sustainable development goals and scale them for global impact. While the mo the innovation factory, which is part of the summit is to build this innovative, diverse and inclusive community to identify the most innovative startups scallops He has an entrepreneur's using AI to achieve the Sustainable Development Goals. Now, like most of the world has gone visited, the AI for Good Summit has also gone bust. And we are moving forward with weekly online programming allows us to reach even much more audience and people through our transparent, so it's always on audio long. I'm glad to let you know that the code for startups and ideas is still open on our website AI for Good, I try to use it. So please, if there's any startups out there, I do innovators, feel free to go to our website and apply to be on these lightweight pitching session, as opposed to the innovation factory, the innovation factory pitching session, we'll have a lot of startups that are being evaluated by Junior experts based on the BI solution, their solid business model. There's innovative potential, and select the startups like the top the best be involved in various AI for Good Summit and good presented as outcomes of the 2020 AI for Good Summit later this year. Before I introduce the startups and mentors and judges who would act as our evaluators in this pitching session, I would like to ask for our attendees. First of all, I would like to welcome our attendees today. And I would like also to ask them to have a very interactive session and use the chat functionality to chat with the judges dementor, the startups and even some chat to everyone. So please make sure that everyone sending their messages shipping or setting the messages up on the chat to everyone to panelists and attendees, just the panelists and that can everyone see this kind of chats. And before we start, I would like also to get a like a sense of who's in the room and see where everyone's connected from. So I would like to ask all the attendees to actually use the chat functionality to say where they're from which country which city, I was talking to myself. So I'm going to Geneva we can Jan I can see on Cyrus, London, Berlin. Zurich, Roseanne? Yes, go super fast, actually. Oh, okay. So that's a lot of diversity from Portugal, I can see as well, between you Singapore, Hyderabad, India, San Francisco, Mumbai, pretty swell. So Saudi Arabia, I mean, we have a lot of people today for this pitching session. So I mean, I think now we need to go forward with our pitching session. And I would like to introduce all the judges. But before I do that, I do some housekeeping rules about the process today. So we're going to have our startups and each of the structure going to have four minutes strict form is to have the pitch. And there's going to be followed by a q&a between the mentors and the judges and the startup. And after the q&a dimensions, judges will be evaluating startup to a new scorecard is going to be sent to us while we are getting ready with the next the next start. So I would like to introduce today's mentor, and I will start with Gary Stewart. So Gary As the CEO and co founder of the nest, and from 2014 to last year actually, he was director of wire UK, where they invested in more than 185 startups that raise more than 260 million dollars. And were valued at more than 1 billion. But he was also an Associate Professor of Entrepreneurship at IE Business School in Madrid. Gary, thanks very much for being with us today. Thanks for inviting me. Know, it's a great pleasure to have you all the time working with us in various IQ events, some very around like six years, in more than six years, you were the director of wire UK, right? And you were one of these, like managers of the world leading exchanges backed by phone can also it's a huge one. And actually now just last year, you start to be the co founder of your own startup. So how does it feel to be on the startup side? I mean, were there a challenges that you might have overlooked despite your extensive global experience?
Yeah, no, I had my first startup back in 2005. And we sold it later, but I mean, going back to it after all the time, I mean, the Most important thing that I realized is it's really hard. It's really time consuming, and that it's a lot easier to give advice than it is to know how to actually scale a business. So I have like a renewed respect for everyone who's out there struggling and hustling every single day. And I think that the most important lesson that I learned is that being in this game for so long, the network is the key having people that can kind of help you to answer all the questions that you may have, because you can't do it by yourself. But if you have a good network of people, that's really really useful.
Okay, so the network and networking, it's all about the people. This is what you always say. We're gonna have a very quick follow up question. So I mean, your startup now School, which 14 stripe as the master class for founders app, right? And if it's a master class or founder app, what would be the useful key messages that you would like to give to the Ivy's or technically startups to take away what's the most important thing that any founder should have? And keep in mind while scaling their startup,
I think the most important lesson is that it's all about the business model. It doesn't make the technology's really great. In terms of solving a problem, but the hard part is actually turning an idea or a product or technology into a business that can scale. And it might have been the case before that people were willing to throw money behind ideas. But I think that for the next few years, people are going to be a bit more demanding money is going to be a bit more of a scarce resource investors are running under the table in a lot of cases. And they're going to be very selective about the companies that they invest in. So make sure that your business model is pretty clear.
Okay, okay. You're talking to us from from London, right? Yep. But you still mean you worked in different different countries to work in Spain? Also in different countries with different organizations like universities and actually wire UK and Spain before? And just 30 questions come to my mind right now. We're talking about this innovation and budget and budget cuts and all these kind of struggles like startups having Is there a certain trends or patterns that defines a startup based on their geography where they come from which country? Or which continent? Or are they all the same, just innovation and it doesn't matter where they come from. But there's some sort of challenges or opportunities that is linked to specific countries and another country or specific region than another region.
So I think that because I've developed like a really huge network. And I'm American, like, I kind of ran away from America, but now I'm running back to America now that I need money. I think that realizes that like, I'm tapping into the US market and tapping into us investors is really key if you're going to scale the business beyond a certain scale. Um, you know, if you look at London, for example, London, I think that they do like 7 billion a year or something like that, in terms of money invested in startups in the US, it would be a multiple, I think, like times 10, or something like that. So what you realize is that like, there's a lot more money in the US, but then they also require you to be a lot more ambitious. Whereas in other markets, you can kind of get away without talking much about the market size and plan. For how you're actually going to attack the market and capture a decent share of it in the US market, that's fundamental, they're only really kind of interested in you. I think if from the get go, it's clear that there's a huge market, and that you are somehow going to be well positioned to capture a big, big chunk of that market.
So, Okay, thanks. And I think I see some of the attendees are asking you some questions about your business more than the chat. So please feel free to answer this kind of question. And also would like to remind the attendees to send their questions or, I mean, using the q&a or using the chat functionality, and to make sure that it's for all panelists and attendees who can can all see. So thanks, Gary again, and I think we're going forward with our next judge. And she's majan as guy so um, Jenny actually is a serial entrepreneur, and an advocate for women's rights and she's a co founder of women in AI, which is a global nonprofit organization on mission to close the gender gap in AI. She's also an advisor to UNESCO European parliament on the future of tech. gender diversity and inclusion in AI. So thanks so much for Jeff, for being with us today. Thanks for having me. So I'm wondering, you're a co founder of women AI, which is this initiative that I mean, I'm sure a lot of people who work in AI, they heard about it. It's amazing. It's doing a great job. And when we speak about AI, the first thing that comes to my mind, our minds actually is, I mean, this gender bias thing that can be in some of the AI models or big sets. So as one of the most important thing and the gender issue. How do you what do you think about these AI and gender discussions right now? Are we have these going in the right direction, or there's just so much to be done?
Yeah, thanks for the question. So your question about if you're going to the right direction or not. We can't deny the improvement that we had over the past years and many different initiatives that have emerged, especially companies you know, they are more aware of the impact of bias they are more aware of the impact of diversity in their teams in their companies, their board members, and these are I see them as advancement. But in terms of actions and concrete results, I believe that we are still far from the needed needed results that we want to have. And for instance, maybe I can give you the example of what are, for example, the AI products or solutions that we're using, for example, for facial recognition, we saw so many great movements, for example, from the companies, giant tech companies such as Microsoft, IBM, Amazon, that they actually started to put a pause on their facial recognition, applications to sort of be sure that it is used for the good and for the good of the society. For example, IBM totally, you know, withdraw their facial recognition, activities and so one, though, so these are I believe, that are good are good impacts, but we still have a huge lack in the, in the talents market, meaning that in order to reach the diversity and the parity, we need to have more talents, and more, not only women, but people from different color, different backgrounds, different religions to get into the tech area and being those people who are developing these solutions. So this is what we're really missing. I think that awareness has emerged. That's a really good investment. Now, we really need to train these people and motivate them in let's say, job offers that we have to be more neutral in the way that we hire people to be more neutral and inclusive in order to solve the issue. So I'm positive but I believe that we should not you know, wait, we should do more.
And it's a good point to mention about the talents and innovation talents it's for, to be diverse and to not only agenda an agenda for sure, but also different industry different nations from different regions. I mean, what else are you on? foreigner and we have worked in AI and IoT, and like in different five different continents as you, as you mentioned, and basically experience what do you think I had to ask this question? What do you think about the COVID-19 pandemic, affecting the tech based innovation and even in the talent aspect or other aspect? innovation? I mean, some people see positively because they say, okay, people now know how much we need innovation. And it's extremely important now to do things differently. And some people see it negatively, or, I mean, for obvious reasons, like restricted travels or budget cuts. So how do you see this effect affecting the innovation ecosystem?
Sure, I met so I see this pandemic as a global challenge crisis for everyone. Of course, you know, tech companies close to the other companies have been impacted but some tech companies that they had an online business so they actually had a growth in the business. So that means that you know, they they could use this crisis as an opportunity. When it comes to startups, for instance, it's very tricky for them because in order to survive, you have so many things that happens to you. So you need to, you know, survive and have cash and, you know, keep up, you know, with your, with your business. So when a pandemic happens across happens, either you're fast enough to pivot, or you actually, you know, need to start, start over again. So it's, it's the truth. And but I saw so many projects they managed to answer to this crisis in a really positive way. And they had, like, multiply their revenues. I believe that the patent they make also I had a chance to be on it. Alliance called Green tech Alliance launch a few weeks ago and we were talking about this pandemic gave us a big chance to actually have a pause and see what could we do in terms of for example, things like climate change, we saw that cities like Delhi or attacking there were so you know, clean, the sky was blue and everything was like so different and We I think we had some sort of an eye opening moment that if we want to do things, we can really do it. So Panam was maybe an opportunity for us to see that things we can do it that we were not really aware of it.
Yeah, yeah, this is a great global challenge, but also have another. Okay, thanks very much Magellan. We're happy to have you today with us as a mentor and judge and move forward with our next judge, which is Joseph Hopkins. So he's actually the founder and the senior managing partner of stitch emerge fund and he's a thought leader in AI syndications security technologies just also leads an innovative emerging technology firm that mainly focusing on IP intellectual property and condition more than the author key Payton's another work in network security identification conference kick yourself. Thank you very much for being with us today.
Your theory again I met you accountant is always
always good to see you and really happy to have your contribution today first in the picture. So does it work in this video interesting field of intellectual property. I mean, of course, the overlap between that. And there's still innovation, there's the technology and you need an innovative emerging technologies firm that has this innovative IP incubation model. So do you think that most of the startups and think B's in AI really start to think about it, think about the texture property when it comes to innovation?
You know, it's great question. And unfortunately, I'm going to say no, you know, I just, I just grabbed a new client yesterday. And I'm sure we're all aware of this controversy that's occurring now with government surveillance when going, you know, given COVID governments are requiring people not to wear a mask, and of course, they're doing lots of surveillance to ensure compliance there, depending on where you are, that's raising lots of controversy. And as you know, one of the areas where AI has really heavily contributed to this is this area of image recognition. And then of course, you know, because because you're using image recognition technology has actually enabled people that are providing service And services to actually do a much better job. But of course, as you know, this has also led to concerns around social injustice. And so I was meeting with this entrepreneur yesterday, and he was mentioning how he has this wonderful idea around how he has some AI technology that can literally, it's designed to police AI to ensure that AI is not facilitating racial and justice and other types of social injustice. It's unbelievable. I do I never thought I mentioned AI policing itself. And, and so I quickly asked him as I would do anyone in business, I said, Well, are you thinking about doing a patent or do you have a patent or etc, he says, oh, and of course I was, you know, a little shocked because this is a really creative idea. Now, of course, at the time, when we finished that conversation, he I put him immediately in contact with my IP attorney, and then we got him on the right track. And I think the reason he basically Did not pursue a patent is because of two reasons. One, he's very, very busy. So he's focused on word design and ultimate trying to get the technology fully developed. But frankly, he actually was under estimating the importance and creativity of his idea. He actually did not believe his idea is important enough to actually get a pen. And, and I spoke to my patent attorney yesterday, and one of the guys I've referred him to, and Paul believes this is a phenomenal idea. And he Paul believes this is going to be easily patentable. So I've got a discussion coming up to him later on today. So I think we've got some good news for him not that he should pursue a pen. It's a great idea. And I think entrepreneurs should do that as well. They should be very thoughtful about their, their ideas, their thoughts, and they should pursue these patents because if they don't, someone else will.
Oh, okay. That's that's a great practical example, just just mentioned now, and this would mean to question that, if you say no, that most of the startups and many startups actually work in the tech domain, they don't think about IP The first thing to work on, I would say why is that is many costs reason like they didn't want to pay or they want to go through this process or you think it's not important. So let me ask this first, how important it is to have to have an IP if you have a AI based or tech based startup, like how much this would add value to a solution or product. And second, if this is not the most of the cases, we won't go for that as the first thing. What what are the reasons behind that?
Yeah, if it's great question. So I think the reason a lot of entrepreneurs don't necessarily pursue it is a few things. They're running really fast. It is it can be expensive if you don't necessarily pursue it the correct way. And it's not necessarily perceived as a critical value relative to all the other things you've got going on. IP is really super important. I think there's probably two reasons why I think a high level there's lots of reasons but there's probably two reasons why I think is super important. One is that it allows you to control your idea. If you have a great idea, and you don't control it, someone else will love to take it from you and control it for you. Right? Look at look at the example of Twitter. Everyone believes that Twitter was was basically an idea that was that was that was that was bought by jack Dorsey, the founder and creator of Twitter. But it wasn't, it wasn't him, it was actually another fellow who ultimately was booted out the company, and he did not receive the second reason why you should do IP that is economic benefit. You should be able to control it, but you also should be able to receive economic benefit and economic benefit, by the way is more important than just making money for yourself. It's also gives you the ability as an entrepreneur to actually create jobs and to actually bring value to the world and society. And you're much better able to do that if you have an idea that can create economic benefit that can actually provide value for people that are working for you and people that have anything for your products and services.
Yeah, yeah, maybe Gave sense. Thank you very much, Joseph, stay with us and have all your questions coming to startups, asking them if they have their IP later on in the session. Thanks, Joseph. We're going forward with Aaron Earnshaw. So Aaron is an entrepreneur, startup advisor, mentor, and he has founded multiple technology ventures in USA and India as well. And Iran is also a topic driver at the AI for health Focus Group, which is a focus group that actually was one of the outcomes of AI for Good Summit and Focus Group by ITU in partnership with WHO and he has an extensive experience in machine learning, deep learning, and also the AI for medicine, which is very important today because I think we have couple of startups working in the healthcare sector. Thanks so much for for being realistic.
Yeah, thank you, Amanda. Thank you for inviting me real pleasure to be part of the AI for Good picking session.
Always pleasure to have your contribution to the AI for Good elements and components center. You have found the extent that AI, which is a company that's using the idea sort of global challenges in various domains, like education, healthcare management. And whenever we talk about practical application, I would always ask myself, which are the domains that you think I can really hold the great promise in transforming in being a game changer, of course, in a positive way. So we see a lot of patients and for education, whatever, but is there one domain, one sector we think AI can be? doesn't have that much, or what's your insights on that?
Yeah, so thank you. I'm a great question. Of course, the short answer is a is gonna affect every remain of human, you know, or all of this kind of technology that comes comes along in a generation that affects almost every aspect of human activity. Now, when we found an extended AI, we had the rather lofty ambition of you know, using AI To solve global challenges, and of course, one of the motivations that I named it extended AI was, you know, extending human capabilities. So I see we see AI as a way to augment or, you know, extend and amplify what we can do with our capabilities. So but when we when I started looking at it, one of the first domains that we got into was health because a couple of reasons for that, I was already involved in a couple of different areas and health and digital health, we ran a website called mad India, whose mission is to provide everyone with better information and tools for better health. So we were already looking at a variety of health problems and when we saw the amount of the kind of challenges that you can solve in the in health with AI, especially with things like you know, medical diagnostics, and the fact that India and country and regions in Africa have very few doctors in certain areas. has a real potential to make a massive difference in those areas. So for example, the application that that we've developed for detecting diabetic neuropathy when I started looking at the numbers, it's pretty incredible. Like there's countries in Africa where you have a country with 20 million people with hardly less than 30 or 40 ophthalmologists. So you, you there is no way to bridge the skill gap. We need AI, which is why of course, you know, it's an integral part of the UN Sustainable Development Goals. You know, by 2030, we hope to bridge some of these gaps and AI is being regarded as the key enabler, excellent way to you know, achieve that right. But the other some of the other areas where I think AI has a massive potential for impact is in energy, optimizing the usage of energy. So we're looking at that as another area education, personalizing education and making it accessible to everybody. Adapting To the kind of learning rates that different people have, that's another area where I can have a huge impact, you know, across all of the 17 SDGs. AI has, you know, a huge potential and of course, Andrew Yang said it best that, you know, he, you know, just like electricity costs AI, the new electricity says that just like 100 years back electricity affected almost every industrial transform, every industry is going to do the same thing. And that's a good way to look at it.
So that's how that's what we excited about.
So that's the same question as going as well, because I mean, you have found the grown up multiple technology engine USA in India. So the same question is here, the AI for Good Summit, AI for Good Summit, innovation factory? We have a lot of ventures with startups come from all over the world. And have you have you ever seen like a big difference between startups coming from maybe a more developed or less developed markets or country For, it doesn't really matter. I mean, are there any patterns or trends to be spotted evolved into geography or divisions of division everywhere?
Yeah, I mean,
that's a good question.
You know, I've been fortunate, as you said, to have been born and brought up an NBA, we have our base in India. So you know, the ability to form the company and start companies in India, because the cup is in India, as well as in the US. And so I've seen both sides. And so to answer your question, specifically, with respect to India and US, I can compare those two, I would say to the US is, of course, one of the most mature most advanced startup ecosystems in the world with Silicon Valley being the pinnacle. Right. But that has been kind of a model that India has tried to follow. And of course, 20 years backwards yesterday, there's a huge difference in India. I mean, even if you go back five or 10 years, you didn't see that kind of mentality in India to create startups because most people would take up jobs but I think in the last five or 10 years, that has changed, because because seeing this huge successes in Silicon Valley and technology is basically permeated all of all the rest of the world right. The other factor which is driven startup growth in India has been the penetration of internet mobile phones, which has made India and now it'll be out almost five to 600 million mobile phone users in India, so almost everybody has access to the internet and to information so, there is a whole ecosystem of startups now happening in places like Bangalore and Chennai and, and and as well as in Delhi. And all of them are trying to follow the Silicon Valley model, you know, trying to create lots of innovative ideas and a lot of great stuff happening. I hear that India is the third largest ecosystem of startups, we've got 50,000 startups, there are some really great you know, any, you know, various companies which are come out of India, but so, there's a Lots of lots of differences in similarities. There's a lot of talent in India. But I think the primary driver for the primary difference within India in the US, I would say is access to capital and resources in
terms of scale, right. So, but I see that in the next 10 or 2010 to 15 years India is going to emerge as a as a big hub for startups. The next big 10 $20 billion startups are going to come out of India. And
yeah, we're hoping that the AI for So thanks very much. And I think I mean, without further ado, we, I mean, I'd like to thank all the mentors and judges and for yourself. So as mentioned, we're going to have a shooting at five startups which you can see now from from the screen having some of them really connected. So we're going to have Germany and interlab strong decision America and have mended from You see, when have also by training from USA now. from the UK and I think that start with other from Germany. So please open your mic and video please. Another good to see you. Okay, good to see you. Good to be here. So Alexandra, as you know your you have your I mean, it's your time to get on but you have you have four minutes, I think the time is gonna start when you start speaking, get yourself ready, share the screen, share screen representation. And afterwards, you can have q&a with the judges and mentors. And and yeah, best of luck, please. Right, perfect.
So first of all, you can see my screen now right the presentation is visible to all of you. I can obviously go ahead. So I will go ahead now. So my name is Alexander stanovich. I'm an enterpreneur in digital media and deep tech since maybe 98 actually counting all to get three m&a exits to global corporations and in 2014 afterwards Missing, really the senseless death of many, many people by myself, and who only died because easily available information was not made available in the situation. My co founder, Sasha qnap. And I decided to put all of our energy resources and also hard into this one ambition and goal, which is to make technology useful for society and for humans. So we set off to deploy cutting edge data technology and AI for a safer, more secure world. And with that, we created Eva, which is an artificial intelligence and data fusion platform for security, safety at risk intelligence, but in the physical world, so we're not a cyber technology company, but we focus on things that actually happen in the real world, so to say. So we monitor and manage data on incidents, accidents and critical events. We structure them and contextualize them to this continuously increasing stream of data So that we can detect patterns and anomalies and send out alerts and predict critical events in the physical world. For that we integrate a lot of different various data sources. For example, social media share data from public authorities, IoT sensors and cameras facts, but also the feeds from broadcasting media and news agencies and professional intelligence providers are available on a VA plus our customers, the users that use our systems, populate a VA with their own data and use AI as an incident management tool. And they can also collaborate like with other customers using their own data, making it available and share all of this we do with the highest ethical standards and with full adherence to the strictest regulations imaginable to privacy data and AI. So we're really like perfectly in shape with that. Maybe for that reason, some refer to us as the European answer to palantir. But others reference to us more than more to better forecasts if not for wind, rain, and sun from things like crime, fire accidents, catastrophes, social and political risks, health risk, of course. So COVID-19 was a very important topic for us in the in the last month and terrorism. Today, we are already operational. So this is not a vision we are real and existing. For example, one and city hall the Metropolitan Police uses our systems to increase the security and safety of their citizens. We are working for public transport operators around the world to support them to keep their operations secure and resilient continues and we support very large organizations to manage like the global security one example as you can see is the HL but other multinationals as well particularly in the case of cope 19 this was important you will find us in situation and comm rooms and command control rooms you will find a event the field mobile applications will call goal like to do something As we available for everybody, it's still like on our plan. And we're working towards that. And in general, you can really imagine it as being like universal data platform for security. So maybe it's sometime in the future, your car will tell you where it's safe to park or to navigate, or your insurance company will use our data quantifies all kinds of organizations and use cases that are relevant. All together, we are soon to be 50 awesome people in Berlin beside and in London, with exceptional skills and everybody really with the hearts at the right spot. And well, together with everybody from the AI ecosystem, at least those people that we like and love to push forward we really work to achieve common goals, which are, of course also Sustainable Development Goals. So we are very aware like a lot of the work that a lot of you guys are doing and happy to be part of this global community to drive the use of artificial intelligence for good. Thank you so much.
Thank you very much. That was also perfect timing you have to say, on the on the minute. So I asked for our judges and mentors. I mean, if you have any questions, they just come in even the videos. I can see guys here again, we'd like to grab these.
Yeah, sure. So thank you for the presentation. Can you tell us a bit more about the number of customers that you have currently? What sort of revenue figures you have? And really, who's the kind of typical user and what are they using it for? And how much do they pay?
Okay, so these are a couple. So right now we're working, I would say like really customers, not to mention like pilot projects and things that are going on six large customers, some of them most known like to the world. The usual application areas currently are for professional security experts and in corporations. For example, imagine the global security situation center of DHL Western, which has to take like account and responsibility like for the security of this Very large organization of analysts. They're working every day to figure out like, what is going on in the world? And where do we need to take action. This can be piracy of the Gulf of Kenya, this can be border closing somewhere in the world for whatever reason, and so far, usually, like such analysts would use emails and outlook and traditional tools to manage this information. And through a VA, they actually can narrow down the wealth of information to only the things that matters to the individual analyst. And
just then other people have questions. So how much do they pay? And do you have any IP?
Yes, we do have IP. So aside of keeping the couple of our technologies a trade secret, we also applied for patterns both in the US and China in Israel and many other places and really manage like a lot of growing IP portfolio. The usual the usual relationship that we currently have with our customers is through licenses. So our customers Make use of the of the platform by using the interfaces like the web interface, the mobile interface. And there are like monthly payments for on a user by user basis. This is a little bit different if customers only want to access data to use it in their internal processes or internal structures, then the pricing space on a per thousand API calls faces. So yeah, like it's the similar way as, for example, Google Maps would, would, would, would charge for the use of their API's.
As Gary Magellan Mojo, go ahead, your question?
Yes, I would like to ask you how many women do you have in your team?
Oh, good question. I think we currently have a might be wrong, but it's eight or nine women. Yeah.
Yeah, might be wrong because because we have we're also distributed already now even though like we are assumed to be 50 people were like three places. I'm in Berlin. In Berlin, we have four people like the majority of our team, our engineers and r&d people, we only started to become commercial end of 2019. So our organization, particularly in Berlin, and London are growing but our r&d organization and now this is the biggest one and I think that should be like people it eight or nine, yes, by referring to what you said before. There is definitely still a lack in, in, in talent coming from universities and academic field, which are also women, particularly in Serbia, where we do have like our RND it is still like a very, very male driven domain, whereby we see this changing a lot and we encouraged this because particularly with what we do as a company and technology, question of diversity on all fields, not just gender related, but also a culture related and And all the aspects of being a human are extremely important because there's so much of subjectiveness in the question of security, safety. Not things that are secure and safe to me do not have to be secure and safe for you, not just based on who we are as people and how we were born, but also where we grew up, grew up, etc. So building a technology that that it will work for the world and this aspects needs to reflect the perspective of the world. So it's definitely something that we encourage and Yeah. Thank you.
Thank you very much. I think Aaron had a question as well. So please go ahead. And then maybe have a seniors also might have a question. So you can go next after
your microphone is muted. I think
you're on mute IRA. All right. Thanks. Yeah. Thanks. Thanks for the great presentation and Two questions, actually. First question is regarding data privacy and regulatory aspects of what you're doing. Since that varies from geography to geography and country to country, are you guys trying to like comply with the local regulations in each of these regions that you operate in, for example, you know, is greater in Europe versus let's say, India or you know, other countries? And the second question is about the concern of consumers, you know, and the and the people who are collecting the data about you know, there's a whole lot of concerns, you know, recently about privacy and data and consumers are worried about surveillance and all of these technologies, which is regarded as invasive to their privacy and how are you how are you how are we addressing those concerns? etc. So
this question can be probably like even answered like in one context. First of all, as you said before, like there are many differences like on whether you do business in different countries in regards of regulation, maybe we are just lazy. So we pick the countries of the areas with the most strictest regulation and apply this like everybody else. It makes management a little bit easier, even though one might say, by having less regulation, you can do more effectively turns out creativity starts with a no. So if you restrict yourself by, for example, saying we don't want to have individual names on our systems, if that is, you know, like a challenge that you that you put like also to rd and to designers, then they will come up with new creative solutions that nobody really thought about before. So it's possible to create something like what we do without actually tapping like into the fields, where we're actually we would have to question like our ethical framework, that doesn't really make a difference in regards of like how it is perceived. Right, like the question of course is, oh, this is a technology company. Most probably will be driven by economic rewards. So they will probably do anything that they can do in order to increase the customer base, blah, blah. So now like the conversation around, like, the concern of users has to be definitely addressed differently than through legal or like to technological aspects, it's really about being very transparent. So we are not hiding, if you come to visit us in Berlin, or any other place, you will see, you know, like we are opening citizens can enter our offices, and we speak with them, give them a presentation over what we are doing, and try to be really precise, over over, over over the, over the ethical frameworks and and the, and the philosophy that drives this idea and what we're pushing forward. And inside of that, maybe at the last node, we started while I'd like to commission in Europe was working on GDPR. So I would consider as being almost like a post GDPR company, so we never had to think old ways in a new way. We simply started with all of the restrictions that we've been They're early enough to be now a little bit ahead of them ahead of most but on the other side, I don't consider the regulations and, and the restrictions that we see really an issue if you want to solve a problem, you can solve the problem even by accepting what is acceptable and whatnot.
Hey, thank you. Yes.
Okay. Very good. I Xander a nice job. It's quite relevant to the very topic I brought up via my introductions. Yeah. So yeah, nice job. I think this is a this is a great business. I think it'll it'll only get even more needed over time. I think man, it's just their customers, I think. No, it's Yeah, I understand the nature of I think of how things are. I just want to follow a little bit on what Gary asked, just qualify. Again, if you don't mind. Are you pre revenue or are you revenue. So you've mentioned you
All revenue. Yes. Yeah.
Okay. Congratulations. Congratulation that good?
Well, I'm the, as I said before, like in 29, late 2019. Sorry, we started like commercializing the organization, and actually booking revenues. So COVID-19 out, unfortunately, it was actually had a large impact on our business and also in that regard that the plans that we have like for the complete year of 2020, we achieved revenue wise in the first quarter already. So the company is in it's in a really good shape in terms of proving what it's capable of. But the problems that we now have to face are not really anymore The proving of whether it's a business and whether it works. It is now really about ensuring that we have the structure and the organizational capacity to actually scale this beyond the level where we are right now. So serving 567 customers is one thing, but serving 20 3050 hundred customers was such a critical aspect is the real challenge still lies ahead. We are who we are we don't we don't have us if we would assume we would could relax now in the upcoming years. No.
Okay. Okay. Very good. And then one last question I can get. Congratulations. I think this is a tough economic environment. And sounds like you guys are really on the right track. Really the last question I know you've got good. Moving on to the next presenter is around your business model. Looks like you have mobile apps. So are you are you attempting to sell to both consumer as well as to enterprise and enterprise? Is that more of like an OEM
So we we chose four separate for different regions in the world, different different strategies. So I would say like in the developed world, saying Europe, us, we don't think it would be wise for us to come up like with an app. It's much more To go like to be embedded like into existing ecosystems so that we don't have to compete on the App Store, like with Angry Birds and Fruit Ninja. Okay, whereby I think like in other regions of the world, we might we might come up like with a with a mobile app for end users. The mobile app that we are deploying right now are for security field forces and professionals. So for police officers or security staff that actually need to go somewhere need to be in constant contact with the security situation center. So this mobile apps rather also like for for governments or for business customers,
okay. Okay. Great. So it sounds like you're you're you're focused on businesses, enterprises, large, small and more OEM system integration. Thank you.
Nice job. Thank you. Anyone, Joseph Morgan, you have a you have a very quick question, please. Because I'm sorry. Just a quick one. Yeah. I just
wanted to because I mentioned the police office and everything and we'd heard about the recent basis and towards the tours like, like, you know, we we heard about it and the application Using data can, you know, sometimes augment the biases we have? I just was curious that are you working with police and authorities? Are you selling it to them? And if yes, how do you manage to, to, you know, make sure that this
so we worked with governments and with police. So we, we, on the other side, do not do everything that a potential customer or customer wants from us. So, whenever we talk about things that are related like to, you know, identifying individuals understanding organizational networks, or for example, street gangs, you know, who's responsible for whom and there are many, many different companies working in this loop for us, it's really more about the fabric of understanding information making information flow. So the decision how a security department of a corporation, the mayor of the city, or the individual police officer is reacting. It's not something that we You're driving through a VA we are providing information hopefully, this information will help to make better decisions, but the actual procedures the actual the actual perspectives and and that that that the individual using our system is having we cannot necessarily influence and I think we would have it very hard if we as a technology company would try to, to go beyond giving advice and sharing our rather humanistic perspective on these kind of things. So, we say you know, like data is something that that should be that should be supporting all of society, not just specific groups. And we all have a very, very long way to go to ensure that, that the power is not not used in the wrong way, and maybe through democratizing part of this power because even technology was responsible for us even becoming aware of What happened in the US, for example, with Mr. Floyd? We probably would never have heard that if there weren't smartphones if people were haven't recorded that and put it back on, on social media. So this has two sides of that equation. And we are very aware of it, but but we will need help of everybody like fighting like for the same cost, the technology will be used for good and not for bad. I hope this answers as short as possible because we did this probably like a discussion for a long evening with a lot of white wine.
Thank you very much. Thanks for just Thanks, everyone. I think we should move forward with our next pitch. And it's going to be very much centered. So it's going to be into labs from the US rain center labs are already there. Please enable your mic and video. Yes, can see the screen being shared very visible. I can also see you so please go ahead. And before minutes and then followed by q&a also like to remind the judges and mentors to use the scorecards for the previous pitch to score our thinking much, and the tone was gonna start with you there when you start.
Thank you, everyone. Hi, good morning and good afternoon, wherever you are, and they're Indra, co founder for Intel labs and we are doing quality assessment for fresh produce in a scalable non destructive and low cost manner using headphones as the quality assessment tool is we all know that more than a third of fresh produce is lost today. Whatever we are doing, only two thirds of that is reaching has been consumed. And on the on the flip side, more than 40% of the population does not have food available to them. Now in such a market, we would assume that the growers would be getting high price but they only get only 10 cents on the dollar which is being sold. A lot of governments have taken efforts to set up prices where they are set ensuring that the growers get a better price for their produce, but it still is based on a fair average quality. There is no premium which is paid to any grower across the world for prices. That's primarily because the whole quality management system of the quality assessment system is manual today, the biggest of the food organizations I've got a person who would visually go through the lot and then give his assessment saying this is good or This is bad. You can argue saying that there are processes and there are metrics against which things are against you still not as good, but if the matrix is like the meters should be uniformly red in color, how do you bring in any objectivity into it? What you may perceive as a uniform red color would be very different from what I perceive as a uniform red color and definitely very different from what I have my foot perceive as a uniform red. Right? So in such a case it is very difficult for any organization any comment a premium for quality based and which is why a lot of farms and farmers are moving away from farming because No matter how much they spend, and they're all sweating on the street at regular market prices. So this is where our solution comes in, it is very simple, we use your hand for capture an image of the sample, which is primarily fresh produce, it could be played out in any way you didn't play out in a particular manner on a tray or any ways it could be in a tote in a crate in a box and a gift whichever way. And then our auditor would read that image. It would then give you the qualities of the sample, not only it would give you how many are good, how many are bad, which is classification integrate a great index, but it would also give you a each individual defect which you can find out. And then all that data is collected and presented as a quality dashboard which can be used to mine actions, insights and transforms. All this is backed by verifiable images and defect logs. For example, you could see here that in the whole image, you can see how many approves somebody that became right shoulder and when you are looking at that data over a period of time, you can start Finding out trends that What is wrong in my supply chain, what are the defects and where are those defects getting introduced? And how can I move ahead? using this technology, we've helped our clients which are primarily food organizations like retailers, traders, aggregators, distributors reduce strong purchases were about 15% essentially, it means that they were buying produce, which was not meeting the quality criteria, which was leading to losses which was leading to rejection, which was leading to dump across and then those have been reduced and now that and also the dump, which they used to throw it and the debut had been reduced to about 30%. So that's a that static saving not only to them, but also saving of food produce which can be used across a passion that there are a couple of other key impacts, which he said. The first one is we've heard a lot of our clients to start giving premium for pricing, for example, in a file we've introduced on our online auctions of cardamom, we Introduce quality criteria close to 70% of India's Cardamom is our guest through our systems. And we've seen that for the same comparable lot today from growers are getting good 30% of growers are getting about 18% higher price realization than then what they were getting about a year ago primarily because of quality being digital and now there are bits which can happen across the place and the buyer is now more
Similarly for one of the very wages, their accuracy of assessment was around 70%. And then today we're using our system it is up to 92% and they are now able to pay better premiums to the growers and give them incentives to bring in better berries matured berries which
has to wrap up quickly because of the time please.
Sure. Similarly, we help the Indian retailers start premium pricing from the growers. The team we've got a team which has got experience in data science agree and vision, computer vision and We are the biggest company which has got the highest database for fresh produce. We are the only ones which are looking at 45 different commodities. And we will funded by various VC partners, NV this Alaskan.
Thank you. Thank you, Dave, I think you can have more efficient during the q&a so that that's fine, but just didn't have the time. So mentors, any any journey anyone wants to ask? Yeah, we have more general reading. So please go ahead.
Yeah, thank you. Thanks for your presentation. I think I'm not sure quite if I got, who is your main customer? Is it your the mass purchasers like the big retailers? Or are they people or people who are using your app at this shop? I didn't understand it.
Great. Primarily, we are in a b2b kind of environment where all the food organizations like I said, retailers grow retailers, aggregators, distributors, traders. They all are consumers who are doing quality checks on them. muster.
Okay, so it's basically a professional app. Is that
right? Yes. Okay.
Okay. Thank you.
I think so my question is a follow up, which is, how much does it cost your average customer? And how long does it take for them to pay back? You know, the cost? Because I'm assuming that like right now, you said there's a low margin, they can just get people to do it manually, very cheaply. And I'm not sure how much it costs and when they see returns, so can you maybe walk us through that?
Right, so we are essentially charging on a service versus the number of commodities number of installations, there's a price monthly price for that. Like I said, we've helped organizations reduce their dump by about 30%. On an average, let's say a retailer has got about 6% dump, which is down by 2%. So that's a straight 2% added to their bottom line. So if you're buying 100 kgs, you're throwing away six kgs. And now you're saving two pieces of that. So that's the kind of ROI VC we've seen returns on these
just Because I don't really understand like fruit and stuff like that, if you get money, that will be great. So how much money? And how much? How long does it take for me to kind of pay this back based on the savings?
Right? So in terms of money, it is anywhere between 100 to $200 per commodity per one kind of a basis in that range. And in terms of returns, we've seen clients realizing returns as early as 10 to 12 months.
Sorry, last follow up and how many clients do you have and how much revenue they
generated? Right. So today we are working with close to 12 or 12 different clients which are retailers, growers and traders. And in terms of money, we we close to an error of about $1 million.
Hi, Dave. Great presentation. I really like the use case especially you know, it's very well aligned with the SDGs in some way, it can reduce food wastage. But I have a couple of questions. So one is, are you doing this at each point in the supply chain? Or are you just doing it at one point? Or, you know, multiple points? Right. And the other one is, of course, its impact on jobs at the grower level. How do you see that? How do you see your solution impacting jobs? And also, one last question is, there's a lot of competition in this space. From what I know, there's quite a few companies. I think, in fact, one of the earlier pitch sessions that was not the company, I'm not mistaken, which is based in the US, which is seems to be in the same space. So what's your main differentiation in terms of your product? And you know, why should somebody buy your product versus you know, somebody else's? Okay.
The first the first one was around the supply chain. So we had, we've been doing this about two and a half years now. We primarily started with the consumer facing organization because they were the ones who are facing high costs of quality and today, it's being To use across the supply chain, let's say for a retailer extreme use at their warehouses at their stores as well as their purchasing centers. This is being used by the aggregators to supply to the retailers and this is also being used by the traders who are purchasing in wholesale. So not today, the solution is being applied across the supply chain. In terms of jobs. It's not about taking somebody so it's not reducing jobs, but it's about enabling the whole quality system with a tool which helps them take the quality in an objective manner. Right. For example, if you look at it, if I am a quality Inspector, I have I have to look at a truck full of truckload of berries which has come in which is what about 50 boxes, each box has about 100 clam shells into it as per the standard sampling after looking at about 6070 boxes into it, which is not possible. So I look at only about four or five different boxes. Each box or each crate would take me about 10 to 15 minutes. I'm reducing that time from 15 minutes to about three minutes now. So I'm helping them do a bit do this job. Mostly. In an objective manner, and instead of writing reports, all the data is getting automatically captured. So it's for a tune, which is helping them do their job in an objective manner. Capture the data as well as ensure that the data is captured at the lowest level and is being used at a noise level so that people can take more informed decisions to still require the quality to team to data informed decisions until it's finally on to the party. In terms of competition. I think we are the only ones which is focusing on fresh produce, we've got a better solution in terms of 45 different commodities. And we are the only ones which are using headphones as one of the primary means. A lot of competition is using different machines, boxes, which what it does is it standardizes your image sticking condition it standardizes the lighting it standardizes it standardizes all the other external parameters, which makes it very difficult to scale. I mean, if you've got a 40,000 bucks, it's very difficult to scale. applied at 100 different procurement points.
I have less than three seconds for the questions. I think we can go quickly with Joseph and then go back to module one because the gentleman asked, so just ask a question, then go back with us.
Okay, very good. Nice presentation. Just a couple of quick questions. When did you start your business? And then secondly, what is your target, targeted geographical reach? So like to get a sense in terms of where you doing business today? Give me a
perspective on that. So we started about, we started about two and a half years back. One of our biggest clients is reliance retail, which is which started about two and a half years back that old element of the solution. They've got 1100 stores, and in terms of geographic reach, we started out of India, that's where we are based. And then slowly and then over the last year, we've started working in Southeast Asia, China, and now we've also got a couple of plans in the United States.
Okay, you mentioned I think had 12 customers or so. Right? Yeah. Okay, so a few in the US, the majority of them outside the US, I'd say yes, yes. Very good. Thank you
guys didn't have time to ask a quick question.
Yes, please. Yeah.
I just want to know, what is the precision of your model, the image recognition, how precise is it?
Right. So that's a very interesting question, because like
there is a very difficult ways to compare it with the the human manual creating itself is about 70% accuracy. But if you look at in a benchmark condition with the experts as well as in laboratories, we are at between 92 to 97% accuracy levels, depending upon what kind of commodities we are working with.
Okay, so your precision you say is 93% accuracy.
Yeah, 90 to 97% accuracy for different commodities,
okay. Okay. Thank
you. Very much. Thanks for everyone thinks that very much was great. And with answers to the questions here, I think we'll move forward with our next pitch, which is Mendel from from USP. So I think she featured you connect it. Yes, we can see you and I hope you can hear me speak.
Good. So please feel free to share your presentation and go ahead and we're gonna have the timer on while you start speaking.
is are you able to see my screen? Nope. Just share the screen again, please.
Yes, no, it's working. Just put on the slideshow and please start.
Hi everyone. My name is Shafiq Pooja. I'm a previous co founder and currently the chief of staff at Mendell I'm delighted to be able to present them Dell were founded in 2018. We currently have 50 full time employees. We're based in San Jose, California with a satellite office in Cairo, Egypt. were backed by DCM, which is a blue chip VC based in Silicon Valley with over 4 billion in capital, focused on oncology and were recently featured by Deloitte as one of the six unique case studies for AI powered research. Our team our two co founders is Dr. crinkle wheel with an mg and Dr. Weil saloon with over 15 years of AI experience was previously funded by DARPA multiple times has 15 years of AI experience in the healthcare domain and my esteemed fellow colleagues.
So she forgot to interrupt I think you're not putting it on the slideshow. So we go to the presentation on the slideshow, just make sure that the slideshow screen to be shared with The presentation itself. Sure. Are you able to see that
No, we see the presentation the slides but we don't see the slideshow. So maybe with your with your cursor, try to click on the slideshow button or f5 or make sure that you're sharing the slideshow screen. Sure. Coming in just once. That's
okay. And you better
Yes, no twice. Okay, great. Sorry about that. So
today's options for for data are currently in the medical space, very primitive. So on one side you have structured data, which oftentimes is very superficial where it does not capture certain important and extremely relevant relevant information. For example, things like staging of cancer, and being able to look at things like PDFs and scan documents are often missed when looking solely at structured data. Other options for being able to look through data are normally involving that human abstraction or human curation of data. This is extremely slow, expensive and very difficult to scale. For example, it takes about a million dollars and 27,000 man hours to be able to curate about 50,000 patients. The last option is AI with very low clinical validity. And this requires heavy corrections by humans and experts to a point where human creation is oftentimes seen as actually a better option. What we've done is being able to organize how to be able to structure unstructured data. In the sample here you can see how the different notes all of the different doctor's notes, lab notes, pathology reports, scan documents, we've developed products to be To understand how this unstructured data can be organized and also being able to be developed into a knowledge representation, we've developed some products called retina, redact, and read that I'll discuss in a second. For starters, retina is our proprietary OCR that we've developed. Right now about 60% of patient data is trapped in these scanned records for things like faxes, printed reports, and oftentimes these this data is not resurfaced, being able to be used for medical and clinical decisions. And then those develop the best in class tool for OCR of this medical data off the shelf solutions and some of the state of the art including Google, fail to recognize at least one word and every four words even when trained specifically on medical data. In addition, we've developed redact which is a proprietary D identification tool. Mendell, as we know it is the only company currently In the world that has AI technology capable of passing the HIPAA accuracy threshold, which is a 99% precision, it's able to redact PII even in the middle of the text without compromising the richness of the document. As seen here, you can be able to see that it understands the inference and entailment being able to be associated and the as models of being able to understand what's kind of going into the medical literature and not simply just machine readable. What we've done has been able to create a tool that after we've OCR and de identified the data, we've created a search engine. However, the search engine is not your typical Google kind of keyword search. This is a semantic search capable of understanding what is being asked of it. This leads to analytics ready data, which can be tracked and verified to the source documents of when you put a search and certain patients are populated. You can see exactly and For what evidence and what reasons that was being decided. This kind of leads to the last in the marquee product, which is called clues. Typically being able to aggregate all of the data and being able to use it into an analytics ready data set takes about six months, and you can't really change the endpoints. With what we've been able to do, we've been able to develop an understanding which we can do in a matter of minutes, and be able to change those endpoints for regulatory submissions. If you see this graph, what would take 27,000 hours, we're able to do in a matter of 15 minutes, what was taking hundreds of thousands of dollars, we're able to do for a few cents, providing a 360 degree view of the longitudinal journey of the patient. You can see here for a visual depiction of how our medical text recognition is vastly superior. This is compared to Google, which right now is considered one of the state of the art on the market, and how much more words and texts We pick up, which allows for our knowledge representation, and the ability to understand this on a larger scale for each record for each patient over millions and billions of different records. Deloitte recently featured us in a case study, as mentioned, and need to wrap up
because Time's up. But
we're integrated a number of different hospitals. And we've recently repurposed our technology outside of oncology, especially given the recent kind of COVID outbreaks and I'm happy to be able to answer and take more questions.
Thank you very much for this and any questions from our judges, mentors? We have Aaron, please go ahead. Followed by Gary.
Hey, good presentation, Shafiq. I have a few questions. I don't think you got into the revenue model, I guess it was. I'd like to understand what your revenue model is. And also, the other thing I was not clear about is I know that we are digitizing all of the paper records And all of that, but how are you doing anything in terms of, I suppose you are doing predictive analytics on top of that, or a just a search engine, which allows clinicians to query the data. Are you doing actual? And are you looking at other areas besides oncology,
sure, on and really apologize for any of the glitches. So first, in terms of our revenue model, we just started generating revenue in 2019. We facilitate and see ourselves as sort of an ingredient basis business. So depending on who our end user is, we're working with pharma, we're working with real world data and real world evidence companies, along with individuals just need to be able to tap into their structured and their OCR and their D identification capabilities. So we have different models, primarily b2b at the moment. But depending on the end user and how much data we can be able to kind of plug in to a specific aspect or take over the entire workflow in regards to, you know, the predictive analytics part hlr health economics and outcome based research, we've been able to move from retrospective decision making, seeing how things work in the past and stitching together, not just 100 patients, but thousands and 10s of thousands to start being able to form things like in silico trials and running virtual trials, looking at more prospective decision making, especially in the act of COVID. By being able to look through this, as you're testing hypotheses and seeing things come through the market, you're having a great way to be able to run a double blind or sometimes a blind study, being able to kind of test these things out in a virtual setting, before being exposed to that in a practical situation.
Just a quick follow up are you doing on a test? Are you also doing images because I saw it was only passively
images so hard. You know, trying to be able to get data is extremely noisy from tables. There's some challenges in there. But our goal is to surface virtually everything coming from both structured and unstructured being able to make That's semantic iqueryable and then extracting those endpoints specifically in providing source document verification. Oh, thank you.
Okay, um, can you tell us a bit more about your IP? And in particular, I guess you said that Google is your big competitor because you're for x better. So how do you defend yourself against Google and why I do this.
So we're patent pending with two proprietary patents. I mentioned it My, my co founder, while saloon, he was funded by DARPA. He's been doing this for about 15 years. In terms of our kind of the way that we've approached this is the we tried to use off the shelf solutions for a lot of our AP. we've integrated with about 15 oncology clinics. And we started with oncology because it's the absolute most complex, and that's why we're now able to repurpose this tech. We've had to build our own OCR, we've had to build our own D identification. And the reason for our D identification is we've got this third party validator which meets and exceeds the HIPAA threshold. So that allows the hospital to share this de identified information with us because they know that it exceeds kind of threshold. HIPAA is extremely important. Google's kind of OCR and they're off the shelf solutions don't meet these thresholds. And that's why we kind of had to refine and build it ourselves, and also build our own ontology, which incorporates millions and millions of records that our algorithms and air models have trained on.
Which I would like to go ahead with one quick question.
Just one quick question. Thank you for your presentation. Could you please explain how do you contribute to the SDGs?
Absolutely. Um, so in the last slide, I put that number 17 essentially the ability for partnerships, democratizing AI as you talk about bias, season eight, I really liked your intro. It's extremely important. To be able to start off at a baseline that there's an inherent bias based upon different physicians across the world. My co founder and CEO, Dr. Cream Kerviel, he started Mendell with the mission of being able to make basically a Jarvis for medicine. And in terms of the kind of initiatives for the UN, this ability to take different data sources. Right now it's extremely biased. And this ability to be able to look from retrospective as we look prospective, you can be able to narrow down this data, geographical specific cities to different countries and be able to kind of make the best decisions. And also we have a team of kind of clinical oncologists that are manually creating and looking through this data to be able to spot any of these blinders on providing for further substantiation.
Thank you, thank you very much. There's if you have a quick one.
Yeah, just real quick and I apologize here. Nice presentation. As I was going to you as as you're going through your presentation, I noticed you've got a ton of capability, a lot of product offerings. So one of the things you want to be careful is not to scatter yourself so thin to where you start to lose focus on the things that you're really good at. Right. So I, I'm curious, you know, what do you see your target targeted to kind of artists, it sounds like your customer sounds like you guys just started to hopefully it sounds like you've just started getting some revenue. So you've got pharma, you got you got r&d departments, Mutual's pharmaceutical companies, you got clean labs. You've got like companies like IMS, you got data analysts, you got lots of different players there. Where do you plan on focusing to ensure because any, any one of those groups I just mentioned, could be a tremendous benefit your company?
Sure. So just have a really appreciate the ability to, you know, caution against going five directions versus going specifically in one and we've been struggling for the last 18 months. The reason we kind of avoided the full commercialization is really understanding our product and our capabilities. So I kind of I demoed a couple and kind of discussed, we do have some other product offerings. But the most important is in a post COVID universe on basically large data trials and the ability to query this data. pharma is not selling because they don't have salespeople going to hospitals, right. And for health economics, oftentimes, you need to be able to demonstrate validity. So for us, the ability to tap into unstructured data is everything with our OCR and our D identification technology. If you miss one of every four words, your entire process is broken when you're looking through only structured data. And oftentimes, there's other competitors in the space that are doing this, but without AI. So if you change cancer, or you change the definition of staging, you can't manually retrospectively abstract that. So there has to be a more sustainable and scalable solution. And we're basically at the helm if you think of the company like flat acquired by Roche, we can do this with a high for a fraction of the speed and time with much more substantiated results.
Thank you. Excellent. Thank you.
Thank you very much. He was really good with answers to all the questions and pleasure to have you so wolf holding our next pitch right now. So our next pitch is by atrium. So by triggering seven, would you please enable your mic and video? And yes, we can see you and please share the screen possible. And once you're on mute.
Just unmute your mic, please. You can hear me.
Okay. Yes. All right. Can you hear me? Yes, clearly and just exactly. You can see the screen as well. Slowing Can you please put in a lighter mode? Yes.
Great. Everyone's full screens.
Yes. dickory look
wonderful. Thanks, everyone. Good afternoon. My name is Ivan divani. I'm the founder and CEO of IO Trillian. We're a little startup looking to advance health care through niche within health care club digital biomarkers. We're developing the neurologic digital biomarker platform called bio engine 40 used to digitally detect developing disease by innovating at the intersection of computer vision and physiology. Were fundamentally reimagining Life Sciences through the principles of many sciences and as a result, we've recruited and built a team with backgrounds across life data and computer sciences. My background includes a receiving a bachelor's and master's bioengineering followed by over a decade in healthcare m&a experience across device and diagnostic technologies before I founded the company, the real catalyst for me founding bio trillion was my own personal calamity with my mother stage four cancer diagnosis which made me wake up and realize that there couldn't be a much better way to detect diseases earlier. So we spend less than 1% of our time in a medical setting, that we're so dependent on it for most basic of our healthcare needs. Consumers advantages access to the 99% observational window outside of the medical setting from what we call the life setting. It's during this much larger window of time called the detection phase, where we need solutions to measure whether a potential diseases develop before becoming symptomatic. As we've seen, COVID-19 is accelerating the urgency of detecting disease signs before becoming sales. Automatic. This has never been more important for both COVID-19 and all time sensitive diseases. The best way for healthcare to be accelerated by the digital revolution is for physiologic measures to move from qualitative to quantitative, then computable and scalable and we believe digital biomarkers and Nish within healthcare holds the solution. So, what if over a billion consumers could be enabled with immediate access to detect health conditions using only the very optical hardware in their smartphone. We've applied the smartphones, camera and other more recently enabling optical hardware to extract extremely subtle and milli scale features in the eyes and face using convolutional neural networks to turn this even digital biomarkers for both cognitive function and neurologic dysfunction. And today, we can now detect drug use and cognitive changes in the brain. Just using someone's iPhone, with our app, of course. Interesting things about AI is that most of it is for focus on the macro world or the microscopic world. But there's this world right in the middle called the Milly world, where features lie that we can see with our own Gen. But those spatial and temporal features can be more accurately quantify it through AI and computer vision. And we believe that that's a massive gap in the market. And those feature sets lie right in the eyes and face. Today, neurologic symptoms can be detected through a pen light test that this is qualitative. You see this on the left side. It's measuring the pupillary reflex, but it's being assessed by the doctors own human visual cortex. This is a really cool qualitative, it's inaccessible, it's expensive, clinician dependent. And scalability is limited for one doctor per every thousand patients, we see a future where it's a one to one relationship with your device and yourself. It's far more scalable, real time affordable. Anyone, anytime, anywhere. And we view pupillary responses to be a next gen vital sign for the neurologic function in the same way you can think of heart rate heart rate as the vital sign for cardiovascular function. So I'm going to play a quick demo. It's about 20 seconds. And it shows basically our subject here is Biggie given a cognitive load a math problem, and we're going to quantitatively measure the synaptic activity in his brain increase as he processes
So he was mentally processing the problem. And you'll see there a spike in his cognitive activity, able to be measured right through his eyes and face. And then you'll see a drop the minute he melt. He answered the question. I'd love to go over this in more detail with you, but in the interest of time, I'll have to go a little faster but by being able to measure the synaptic activity, this can be used for measuring changes in cognition quantitatively. So for example, Alzheimer's or dementia has been giving someone the exact same difficulty stimuli and measuring whether that becomes cognitively more hard to perform that, that that problem. You can see whether the cognitive decline is occurring on a quantitative basis.
Thank you, sir. But actually, we're taught time's up
and Okay, and I'll wrap up. Thank you. Our ministers come from backgrounds across the health and tech communities, such as Nvidia, one medical and others. And our Sustainable Development Goals are aligned with go three and gold 10. Namely around neurologic health anomalies, the scalability of anyone, anytime, anywhere, 3 billion people on a smartphone and be happy to answer questions. Thank you very much.
Thank you, Priscilla. Any questions? Yeah, we have urn, please. Yeah. Hi.
Great, fascinating. presentation, great use of AI. But question is clinical validation, as well as how you're training your models, just to keep it very brief.
Okay, so the training data set is a little unconventional. We don't think Shouldn't is having a data and a label with the label being really the endpoint. And there's a paradox here because we don't have labels with people or AI. And so what we do, label it to the actual people changes, and train our convolutional neural networks to be able to detect the changes that a very subtle level, and we use pathophysiologic causality, to then back from, from those changes to cognitive function and your logic dysfunction, based on what we know, through causality of the brain,
or clinical validation.
Right, so for clinical validation, so we've we've started to test this on a low sample set of friends and family. Were able to detect drugs Alcohol, opioids don't ask how we got and a few other things and, and also changes in cognition. And then our next goal is to run a pilot observational study at about 50 to 75. population size is the next step.
Okay, great. Thank you.
So could you tell us a bit more about kind of accuracy rates or kind of how we know this actually works? And then what would the legal requirements be? Because I'm assuming that if all of a sudden you're kind of giving the sort of health diagnosis and maybe the legal implications, and any IP? Yep.
I'm gonna go to the slide that really assumes a lot of things.
That's the slide.
we look at healthcare through phases, most of its occurring in the interventional phase. Which is typically diagnostics and treatment. Our goal is to be right in the middle in the detection, not to be a diagnostic, but to be pre diagnostic, almost closer to wellness and to catalyze an earlier move into the clinical setting for clinically confirmatory diagnosis. So there's a lower regulatory bar, but there's still validation that's required. On terms of IP, we filed our ninth patent just last week. And
was there anything else that your question?
accuracy, okay. So it depends on which endpoint we're looking at. So for drugs, alcohol, we're able to do about 85% of the time. opioids is around 70%. And then, the cognitive If change, which basically, we mark against the strength of cognitive load. So basically the complexity of, for example, math problem that's in the 70s right now. So but we've we have not gotten asymptotic yet because we have a lot more data to still train. And it just takes a little bit more time to to get to those higher numbers, and they'll start to get asymptotic. And we've used a training data set of over 10,000
images and videos.
Thank you very much. I think it was a good solution. Definitely in touch with you for various uses when you need to do this very much. And I think we're moving forward with Next last one definitely is the startup and reaction time. So now to the last one and casual lens from from the UK. Are you ready? I can see your hungers here. If you speak this audio
Hello everyone, can you see
my screen and hear me properly? We can hear you properly. Can you share your screen again?
Yes, just put them in slideshow mode and you can see it.
Please go ahead. You have four minutes, I'll let you know. And
yeah, I'm representing causal lens. We're deep tech startup founded in 2016. I'm based in London. At causal ends our vision is to predict and optimize the global economy while having a positive impact in our world. muslins helps all kinds of organizations to optimize their processes and reduce waste. For example, we help some logistics companies that want to improve their cargo aircraft capacity factors. And we can do this by simply inputting past cargo data, together with some economic indicators. And then this way that goes on as technology makes autonomous predictions of these cargo demand for each point of interest, continuously updating in real time. This helps organizations optimize their fleet, reducing the number of required cargo aircraft and trucks, leading to large reductions in co2 and other harmful emissions. There are many more ways in which causal ends can help make the world a better place. Our technology can be deployed on the edge. This means that causal ends can be fitted in into hardware that is not connected directly connected to the grid, such as turbines and a wind farm, forming an Internet of Things ecosystem. That enables them to maximize energy production. This can be further coordinated with predictions for electricity demand. To help balance and optimize the electric grid, causal and optimize a Yeah, automate and optimize this entire process, from gathering the raw data all the way to optimizing the business outcomes. However, I think ology goes well beyond automating workflows and processes. We are developing a new generation of AI, which we call causal AI that enables machines to understand cause and effect. Conventional machine learning approaches fail when making forecasts in the real world in dynamic environments. Conventional machine learning models are static, and they do not date as new information arrives. However, these kinds of problems are ubiquitous in our day to day lives. Also Traditional machine learning models are based on historical correlations. So in this way models fit past data, but do not look at the true things that drive your target. causally, I can zero in in these true causal drivers without being fooled by any spurious correlations. And unlike black box machine learning models, causal AI is also inherently explainable. It has the ability to imagine scenarios that have not been encountered in the past. This way, it allows it to simulate counterfactual worlds and does not only have to rely on the usual training data. As a result, causal AI is much more accurate than standard machine, state of the art machine learning methods. In a state we recently did, we saw that the accuracy can be up to 42% higher than the current state of the art. causal links offers endless possibilities beyond energy and logistics, as we've seen, are dynamic causality infused technology can be used for food production. In this way, maximizing agricultural yields in healthcare, improving hospital operations, and optimizing transport networks, among many other applications. Cause lens is run by scientists and engineers, the majority holding a PhD in a quantitative field. We are led by our high achieving founders who have come from successful backgrounds in both academia and industry. And we are supported by a distinguished advisory board, which you can see here, and some top institutional investors. If you'd like to find out more about how causal this technology can be used for good, is it as a causal call. Thank you very much. Happy
thank you very much for the presentation. If you have any questions from our panel, mentors. Yeah, firing Please go ahead.
Yeah. Hi, Alexandra. Question is the causal inference model that you're talking about in AI? Is this a new paradigm? Are you building on top of let's say, convolution, neural networks, or recurrent neural networks? Are you? Are you inventing a new? Looks like you are trying to do something brand new? Are you building on top of what already exists? Or is any academic research to back up what you're trying to? Do? You have any peer reviewed articles or research? I mean, you know, this trying to understand the basis.
No, those are all great questions. So the way we work with this is we look at what there is out there in the open source community. There has been a lot of good research so far in in this concept of causal AI. However, what we have found is that it's it's not quite there yet, right? There's been many Like good advances, but it's nowhere near where we would like it to be. So so we do use some of what's out there. And then we have a research team that builds on this. And yes, we have some, some, not only some some papers that are ready to be published, but also some some patents on the area that that so we do have some some IP, but the patents on this specific thing have not yet been.
Do you have a working model, which actually does these predictions? Are you still in the process of developing it?
No, no, absolutely. We we work with many clients on this and they find that it's been very successful for them. Yeah, and we just tried to keep improving. We know where, as we said, that we've done good steps and proven that we can do better than than what's out there so far, but nowhere near the end of the journey.
Okay. Thank you.
Yeah, so good presentation, but a couple questions about business. Model who pays? How much? And how many customers do you have? And then relatedly, there are a lot of other companies out there trying to, like optimize kind of, you know, cargoes and stuff like that, you know, in lots of different countries, different players. How are you better? Like, what is the Who's your main competition? And how are you going to beat them?
Right? You have many questions. If I forget any, please let me know. I'll get back to it. So we have many customers already in the double digits. And we operate licensing mall, where they pay monthly, it varies a lot in terms of how much they pay, and depending on the size of the company, it can go from from 10s of thousands per year to millions per year. depends on the
size of the company. Sorry, the question
was, well, let me just follow up on that one real quickly as to how much do you guys have in revenue currently, like what's the revenue today?
Yeah, I actually believe Personally, I don't know, but I am not allowed to share that anyway.
Okay. And then the other question is competition. There's a lot of there are a lot of people out there doing the same thing, or trying.
Yeah, sure, absolutely. So, there's a few things that distinguishes above the competition. There's also these big players that are now getting into the field that we, that we work in. And one of the things is that we built the first fully integrated time series platform to do this. And therefore, this is all built from that perspective from scratch. So this is how our clients have told us that we outperform some of the competition, but also we have this different angle, on on causality, which other competitors, as far as we know, are not currently doing yet. So this is how how we distinguish ourselves and why we found that for the problems that we're working with, with perform better, so far,
Thank you much. Well, Johnny have a question for you. This would be the last one for this pitching session. Yeah.
And I will ask the same question how many women do you have in your engineering teams development team?
I believe we have five women in
out of harmony
over 20 Yeah, it's something where we're currently hiring a lot of new reputation and we hope to get many more so please, if you're listening and you're interested, any anyone but also especially any any ladies that would like to apply with more happy to have your applications?
Happy I have
100 Great, good condition. Great answers to questions. And I think by this week, we have reached the end of the Special thank you very much everyone. I mean, we're overtime as usual, but I cannot stop the QA From from having that split. And I would really like to maybe ask very quickly our judges and mentors, and maybe each one of them to give like 30 seconds final economy about this session about any insights they would like to express or anything they felt or I mean listening to these startups or and asking some questions to the start. So maybe I would start with Joseph and Joseph, just like final final remarks from your site.
Thanks for inviting me. Absolute pleasure to meet everyone and nice to see you again. I met I felt that the all of the pitches incredibly helpful. It's amazing. The amount of thought and intelligence that's going that's gone into these ideas, and I'm also very pleased to hear that everyone seems to have IP underway. So congratulations, guys. That's exactly what you should be doing. So good job. There. I would encourage all of you to stay very focused. Try not to spread yourself too thin, if you're in it to make money, focus on that as well. Some of you seem to be a little more academically oriented than others. But I think, on the whole, this is all really good stuff. And, and I think, you know, I met just to close it here, these these companies represent I think, you know, this is a really fine representation relative to what you folks are trying to do at the UN SDG as well as social impact. So congratulations, that as well.
So thank you. Thank you very much for this module. I'm just finding finding remarks and comments. And I have to tell you, because you asked about this actually, we should have had another woman on on the judges judging panel here today, but unfortunately, couldn't make it so we're having a perfectly gender balanced panels all the time.
Thank you so much, man. It was a pleasure to be on this event and very thank you so much all of you, the speakers, the startups a pitch, we just have one comment. And I think that it was kind of general for all the pitches. I saw like all the founders were men and in most of the teams I saw, I think there was five persons participation of women and not many in the engineering. I think that is very important, of course, for startups is most important to at the beginning, survive and have cash and revenue. But it's very important for you and your long term business to have more diversity. It's not only about women, of course, I see that you had a great diversity in terms of skills, but diversity in terms of other aspects such as gender, background race. It's very important and it will help you in the long term. So please, please keep that in mind and of course, it's a higher end To recruit, but there are so many amazing women already out there that are already in the market to be hired. So they exist. We have, for example, I don't know, 100,000 women in tech, we have like 10,000 women in AI communities. So please tap into those communities and talents. And yeah, thank you so much, everybody.
Great. Thank you very much. Gary, would you like to give just a quick 30 seconds for the comments, remarks?
Yeah, no. And a lot of ways, I'm gonna repeat what I said at the beginning, which is a lot of great ideas. But I think that ultimately, you guys just have to remember that judges, whether they be investors or potential clients are seeing lots of companies that are claiming to do the same thing. And what really differentiates you is if you have any traction, or if you have any revenue, well, I guess that's the best form of traction. So in a lot of cases, those were only coming out in the q&a, as if they were kind of incidental where you kind of spent most of the pitch talking about the product. I think that you need to kind of make sure that you understand that the traffic And whether or not you're actually generating revenue and have real clients, that needs to be almost the core of a presentation, not incidental, or something that comes out in the q&a. But good job.
Record. Thank you. Okay. And our last final remarks, please.
Yeah, thank you, man. And thank you to all the presenters, I think we saw some really great ideas and all of them aligned in a way with the SDGs. So it's really nice to see that, and also a lot of diversity not in terms of gender, but in terms of the startups from different regions of the of the world. But my main takeaway from this to the products would be that to focus and, you know, trying to solve one problem in your niche rather than trying to solve a whole broad range because there's some of them which have very broad applications, but in order to be successful and to really generate revenue, as Gary was saying, you need to be focused right on one one area and see if you can get great traction quickly. be better than your competitors, differentiate your product, you know and go after what you're good at. But the last thing is don't listen to what he is saying, Listen to your market and listen to your customers have been bashed The most important thing. I mean, we're just looking at it from our lens. And that may not matter. We don't know your customers, we don't know your market, go with your gut feel go with your strengths. And I think we'll be successful and all the best. And one final thing, all of the health related was means look at AI for Focus Group on AI for health, less applied for whatever you're doing at some beta mission is there so please check it out. Thank you again. Thanks. Very
good. points. Think Thank you, everyone. So I mean, thanks again, everyone, for the judges startups with the pitches and of course AI for health Focus Group, the the Ico and WHO Focus Group, you can just Google it, it's going to be in the first search results for you. And we'd love to be in touch with everyone. And by the way, I just before closing to let you know that to return for next week on Thursday, second of July at 4pm CST time we have a webinar on how to develop accountability solutions. And we also have another webinar on Tuesday seventh of July at 3:30pm CST time as well on our on AI aviation. So they are now being shared on the chat. So please stay tuned in that for our next webinars. Thank you so much everyone. And so if we've gone over time, but it was great dynamic session, and have a great weekend.